Even though there's no visible abuse right now, you know, Google's motto also used to be "don't be evil".
I understand that you have to comply with law enforcement, but actively attacking the users of one of your customer's websites is super rude.
The more of the Internet you've got running through your service, the more appealing a target you are for not only domestic government pressure, but attempts from foreign state actors to compromise the service (through not only hacking, but espionage and blackmail as well).
It's not great.
Also I feel like Raid Forums is a bit mis-characterized in the article. It was largely a forum for people who collect OSINT about breached websites, not really a market place, and in the years that I spent there, I never saw people selling actually carding details, like they claim in the article. I used it regularly for my day job.
It seems therefore entirely plausable that the admin handed they keys to the castle to the FBI anyway, or at least gave Cloudflare the okay to go ahead.
I can't find a shred of evidence that Cloudflare were involved directly in making the phishing page or even complying with the FBI.
We used it at a job I had and it made sense for business continuity reasons. But it is centralizing the internet and they are the gatekeepers. Not a good thing
I also don‘t really get their argument here?
Cloudflare can collect your traffic history, but can only connect that history to your originating IP + timestamp. Their official client may be able to collect more info though. But, warp is just wireguard, so you do not need to run their official client there are shell/python scripts floating around to get the keys / endpoint IPs setup for Warp to use with std. in-kernel wireguard.
Further, all the telcos in the US are known to have colluded in illegal NSA spying on Americans. Cloudflare has not been caught at this yet. So, you can look at it as a choice of exposing your browsing history to an entity that may be not be lying and actually is not snooping vs. telcos that are known to have lied and definitely have and are likely still snooping.
That's exaggerating quite a bit. Maybe in 2005 they had that sort of insight, but with HTTPS everywhere things are different. Your ISP can only see which IPs you're connecting to, possibly which hosts you're looking up depending on your setup but DNS-over-TLS and the like will put a wet blanket on that.
Cloudflare (even without warp) has a much clearer picture of your browsing habits. Not only do they see which webpages you are requesting since they're situated as a MITM between you and a significant chunk of the servers online, they do quite a lot of browser fingerprinting and tracking for bot mitigation that could, theoretically, be used to identify humans as well.
Your point about cloudflare having even more access to your browsing details than the list of sites you have visited that your ISP can collect is a good point. It is kinda crazy how so many companies are OK with a 3rd party terminating TLS for them. And, back on the first point, most sites that do support ESNI today are behind Cloudflare (makes your point even stronger).
But, still, Cloudflare would have to be snooping on content to correlate identity (at Cloudflare scale, that means they would have to already be targeting you), while your ISP already has it.
For me personally (stuck with Verizon which is known to snoop and sell data), I prefer "trusting" Cloudflare until they are shown to be a bad actor like Verizon too.